Download series resistance effects on solar cell measurements

Advanced
Energy
Conversion.
Vol. 3, pp. 455-479.
Pergamon
Press, 1963.
SERIES RESISTANCE
EFFECTS
MEASUREMENTS
MARTIN
Printed
in Great
Britain
ON SOLAR CELL
*
WOLF and HANS RAUSCHENBACHT
Abstract-Current-voltage
characteristics of photovoltaic solar energy converter cells are obtainable by
three methods, which yield different results due to the effects of the cell internal series resistance. The three
resultant characteristics are : (1) the photovoltaic output characteristic, (2) the p-n junction characteristic,
and (3) the rectifier forward characteristic. Choice of the proper method is necessary for obtaining the
correct information for the individual application.
Most frequently used, e.g. for the determination of solar converter performance, is the photovoltaic
output characteristic. A quick way is described for deriving such a characteristic for any light level from a
corresponding characteristic obtained at a different light level. This method involves two translations of the
coordinate. system and requires only the knowledge of the series resistance and the difference in light intensities or short circuit currents.
An inversion of this method permits an easy determination of the series resistance, involving measurements at two arbitrary light levels of unknown magnitude.
The effects of series resistance consist at high light levels in a flattening of the photovoltaic output
characteristic and a related drop in the maximum power point voltage. The resultant decrease in efficiency
has to be overcome by series resistance reduction for solar cell applications with optical concentrators or
for space missions in closer sun-proximity. In cell portions progressively further distant from the contact
strip increasing cell voltages develop, approaching open circuit condition at very high light jntensities. This
yields a reduction of current contribution from those portions of the cell and a deviation from the normal
proportionality between short circuit current and light intensity.
The direct measurability of the p-n junction characteristic at high current densities without series resistance effects by the second method provides a powerful tool to the device development engineer, besides
yielding a second method for the determination of the series resistance. Results from the application of this
method indicate that, in the current density range as used in solar energy conversion, the silicon solar cell
characteristic is much more closely described by the diffusion theory for p-n junctions than was previously
believed.
1. INTRODUCTION
LIKE all other known generators of electrical power, solar cells possesssome internal series
resistance.This internal seriesresistanceis so important as to detemline the current-voltage
characteristic of most of these power generators. This is, however, not the case with the
solar cells. Rather a p-n junction, internally contained in the solar cell, determines the
current-voltage characteristic of the device, with the seriesresistancecontributing only in a
secondary manner. Nevertheless, the internal seriesresistanceis of sufficient importance to
have caught the attention of the device development engineers, which led to its reduction
through decrease of contact resistance and through application of grid-shaped contacts.
Also the power systemsengineer concerned with the application of solar cells has to pay
attention to the cell seriesresistance for the proper evaluation of current-voltage characteristics, for the matching of cells and for the prediction of solar cell output at differing
light intensities. In order to further spread the knowledge about the various effects of the
internal series resistante and about their proper interpretation, a summary of the series
resistance effects on solar cell measurementsand applications is presented in the following
paragraphs.
* This paper was presented at the 1961 Pacific General Meeting of the AIEE, Salt Lake City, Utah,
August 23-25, 1961.
t Heliotek, Division of Textron Electronics, Inc., 12500 Gladstone Avenue, Sylmar, California, U.S.A.
17
455
456
’-.
THE
TIHREE
CURRENT-VOLTAGE
OF SOLAR
CHARACTERISTICS
CELLS
Current-voltage
characteristics for solar cells can and have in the past been obtained by
three different methods.
The most commonly used method applies a fixed illumination. usually of known inten..
sity. and a resistive load which is varied between short circuit and open circuit conditions,
while measuring the voltage across the solar cell terminals and the current out of these
terminals [l]. This method of measurement applies to the solar cell in its normal photovoltaic mode of operation , and the current-voltage
characteristic obtained in this manner
is therefore called the “photovoitaic
output characteristic.”
Figure l(a) top S~IOWS the circuit
diagram for this type of measurement. includin g the generally applied equivalent circuit
diagram for the solar cell [Z--4]. Since this paper is primarily concerned with measurements
for applications of photovoltaic cells at relatively high current densities, effects of shunt
resistance can be neglected and are therefore not even mentioned in most cases discussed.
The following equation is frequently used to describe the current-voltage
characteristic
obtained by this method:
I = 10 i exp +(V-IRS)
- 1 --II,
(1)
r
\
L
I
1
This equation reproduces the obtained characteristics sufficiently well for most cases. Since
a solar cell acts as a generator in this test, the current-voltage
characteristic is obtained in
the fourth quadrant of the current-voltage
plane. The quantity IL, called the light generated
current, is proportional
to the incident light intensity, if the spectral distribution
of the
radiation is not varied. The magnitude of this current is further determined by material and
geometry factors of the solar cell, but is independent of the current-voltage
characteristics.
Other quantities of equation (1) are the terminal current and voltage, I and V respectively,
and the internal series resistance R,; IO is the diode saturation current determined by
material properties; q and k are the electronic charge and the Boltzmann constant, respectively, while T is the absolute temperature and A a dimensionless constant between 1 and 5,
but in most solar cells near 2.5 to 3.
The second method tests the solar cell like a diode without application of any illumination, but by supplying d.c. power from an external bias supply (see Fig. l(a) center). The
current-voltage
characteristic obtained in this manner is therefore called the “diode forward
characteristic”.
Again the voltage across the solar cell terminals and the current into these
terminals are measured. The characteristic
obtained by this method falls into the first
quadrant of the current-voltage
plane. It is described by:
I = IO exp &T(V1
i
IRS)
1
-
1)
(2)
The diode forward
characteristic
differs from the photovoltaic
output characteristic
described by equation (1) by the absence of the light generated current IL and by the resulting positive direction of the terminal current I.
The third method for obtaining current-voltage
characteristics from a solar cell appears
more sophisticated than the previous two. The solar cell is again illuminated, but in this
case with variable light intensity. The amount of the illumination does not have to be
known, if the value of the light generated current IL can be determined. This condition is
fulfilled when the magnitude of the cell series resistance is sufficiently small so that the
output current 1 of the device, when measured by the photovoltaic
output method, is
Series Resistance Effects on Solar Cell Measurements
457
constant for all terminal voltages between 0 and O-1 volts. In this casethe measuring circuit
may consist of a switch, a high resistance voltmeter and a low resistancemilliamperemeter,
arranged as shown in the cirtuit diagram (Fig. l(a) bottom). The voltage drop across the
milliamperemeter should be less than 50 mV and the resistance in the voltmeter circuit
.(I)
I =
lo
/-
SOLAR
{i [
EXP
&V-
IRS)]
-SOLAR
CELL
-I}
-IL
I<0
v>o
CELL
DC
POWER
SUPPLY
I = I,IEXP
[&
b-IRS
/-SOLAR
(i)
r----l--,
I>,0
1]- I}
CELL
vao
1~1‘
b
1
FIG. l(a). Three methods for obtaining current-voltage characteristics on solar cells.
(1) Photovoltaic output (constant illumination).
(2) Diode forward (without illumination).
(3) P-IV junction (variable illumination).
more than 100 kQ The measurement consists of determining the short circuit current, ISC,
which under the presented conditions equals the light generated current IL, and the open
circuit voltage Vo for every light intensity setting. Each pair of corresponding short circuit
current and open circuit voltage values is plotted as one point in the first quadrant of the
current-voltage plane. Through the variation of the light intensity, a successionof such
points is obtained
which presents the desired current-voltage characteristic. A method
very similar to the one described here was first applied by Heeger and Nisbet for the matching of solar cells [SJ, and was again used independently from the investigations described
458
MARTIN
WOLI:
axi
HANS
RAUSCIIENUAC~I
in this paper for similar studies by Queisser [6]. If at higher light intensities the condition of fhtnesS
of the photovo~taic output characteristic near zero voltage is riot fulfilled,
then the light intensity will have to be measured independently, its value expressed in
later
MILLI
200
AMP
---T------
r:
175
i
t--
I
(
.I-SOLAR
CELL
_
1
A-N-
12
----c
FIG.
>’
I
.__.. --i.
TERMINAL
VOLTAGE
l(b). Three current-voltage characteristics obtainable on the same solar cell.
equivalent light generated current, and in this form entered into the graph. The measure-meritof light intensity can be performed by means of a more suitable solar cell, or of other
photometric devices. The current-voltage characteristic obtained by this method is described
by equation (3):
Series Resistance Effects on Solar Cell Measurements
459
It is derived from equation (1) for the photovoltaic output characteristic by setting I = 0
and replacing V by the open circuit voltage VO. In this casethe term containing the internal
series resistance Rs vanishes. The resultant expression is identical to that for the diode
forward characteristic except for the absence of the seriesresistance term.
Figure l(b) shows the current-voltage characteristics obtained on the same solar cell
by applying the three methods described. A translation I’ = I + IL has, however, been
applied to the photovoltaic output characteristic in order to move it into the f&t quadrant
of the current-voltage plane and to make it pass through the origin of this plane, thus
facilitating comparison with the other two characteristics. The observed differences between
the three characteristics are readily understandable from the previous description of the
three methods for their generation. In the photovoltaic output characteristic, a current is
generated internally in the cell, giving rise to a voltage acrossthe p-n junction in the forward
bias direction and causing current flow through this p-n junction. The difference in current
between the light generated current IL and the forward current through the p-n junction
flows through the device terminals, but also results in a voltage drop across the internal
seriesresistance. The measured terminal voltage is therefore smaller than the voltage across
the p-n junction by the amount of this voltage drop. In the diode forward characteristic,
the current is supplied from an external power supply, not using the internal generator.
All of the current passing the solar cell terminals flows through the p-n junction, but it
passesthe internal seriesresistance in the opposite direction than it does in the case of the
photovoltaic output characteristic. Therefore the voltage across the p-n junction is smaller
than the terminal voltage V. In the p-n junction characteristic finally a caseis obtained in
which the effect of the internal series resistance is eliminated. In the open circuit voltage
condition, the terminal voltage is identical to the voltage existing across the p-n junction,
while the short circuit current measuresthe light generated current directly as long as the
voltage drop across the internal series resistance is sufficiently small as to cause only a
negligibly small current to flow through the p-n junction in the forward direction.
The curves of Fig. l(b) substantiate this explanation. The photovoltaic output characteristic exhibits terminal voltages which are smaller than those of the p-n junction characteristic by the voltage drop over the internal seriesresistance causedby the current flowing
through this part of the circuit. Conversely, the diode forward characteristic exhibits
voltage values higher than those of the p-n junction characteristic by the voltage drop
causedby the current flow through the internal seriesresistance.
The foregoing discussionsalso outline the range of application for the different types
of characteristics. Whenever the performance of a solar cell as a photovoltaic energy
converter is concerned, be it for the determination of its power output, for the matching
of different solar cells into a system, or for the extrapolation of the cell output from one
intensity level to a different one, then one has to use the photovoltaic output characteristic.
If the characteristic of the p-n junction itself is to be studied, then the p-n junction characteristic obviously has to be obtained. If the solar cell is to be used as a passive device and
its performance without illumination is to be studied, then the diode forward characteristic
has to be taken.
3. THE EFFECT
OF SERlES
BETWEEN
SHORT-CIRCUIT
RESISTANCE
CURRENT
ON THE RELATIONSHIP
AND LIGHT
INTENSITY
The light generated current IL is proportional to the intensity of the incident radiation
UP to extremely high light intensities. Numerous measurements undertaken in order to
check this relationship appear-cd to show deviations at light intensities orders of magtutude
below those where deviations can be expected. These supposed deviations wet-c caused by
the clfect of the series resistance on the short circuit current which had been 1lleasure-j
instead of the light generated current. Such a substitution can properly be made on ideal
solar cells with zero series resistance, and on real solar cells at sufficiently low light intensities
only. At light intensities high enough so that the product of internal series resistance and
terminal current exceeds 250 rnV, the short circuit current can no longer be considered
identical to the light generated current.
-4G-
-
-----
--
-3
z_
ti
t-
-._--
-
-.
--
.
-18C.
-0.6
-3.5
-,).,l
-3.3
-,>.LJ
0
Terminal
voltage,
0.1
0.2
3.3
3.4
0.5
0.6
-
o-7
v
FIG. 2.
Figure 2 illustrates this fact on the photovoltaic output characteristics
of two different
solar cells taken at five values of solar irradiance between 50 mW cm-? and 400 mW cm-?
One of these solar cells is a gridded solar cell with a series resistance of 0 * 38 Q, while the
other solar cell is a nongridded cell having a series resistance of about 3 - 5 Q. The low-series
resistance cell exhibits rather square characteristics up to 400 mW cm-2 of solar irradiance,
with the terminal current staying constant for variations of the terminal voltage from
0 to 200 mV. Th’ IS means that the short circuit current is identical to the light generated
current throughout this irradiance range. The high-series resistance cell, however, exhibits
successively more rounded characteristics
at increasing light intensities, and the short
circuit current deviates from the light generated current already at.200 mW cm-2 of solar
irradiance. At larger values of solar irradiance, the terminal current does not reach its
maximum value at zero terminal voltage, but continues to increase with increasing negative
terminal voltages, until it finally reaches its maximum value and equals the light generated
_.!
.
*
I
Series Resistance Effects on Solar Cell Measurements
461
I’
I
:?
current. The photovoltaic output characteristic can readily be extended into the third
quadrant of the current-voltage plane, if the resistive load is provided by an external
source which permits the simultaneous application of negative voltages.
:.
‘-
200
R =0*38R
cell A-N-12\
Solar
FIG.
irradiance,
mW
cm-*
3. Short circuit current versus solar irradiance at different internal series resistance values.
Figure 3 presents a plot of the short circuit current as a function of solar irradiance for
the two cells used for the measurements presented in Fig. 2. This plot shows that the low
series resistance cell exhibits a linear relationship between short circuit current and solar
irradiance up to 400 mW cm-2 which is identical to four times the maximum solar irradiance
reached at the earth’s surface at sea level at noon time on a clear day. The graph also shows
the considerable deviation from linearity for the solar cell with the high internal series
resistance.
MARTIN
462
WOLI:
and
HANS
RAUSCIII:NI~ACII
The cause for the deviation of the short circuit current from the light generated current
at high light intensity and with large series resistance, particularly with Iargc diffused layer
sheet resistance, lies in the voltage drop across this series resistance. This voltage drop can
result in an appreciable voltage across the II-H junction in portions of the cell, although
zero voltage condition exists at the terminals.
CONTACT
BACK
STRIP
CONTACT
--,
CELL
0.2
FIG.
4. Diffused
layer
04
voltage
05
08
DISTANCE
versus
IO
I2
14
SIZE
16
IX
IE
2
2.3
cm
cm
S
distance
from
contact
strip,
cell terminals
short
circuited.
This effect is illustrated in Fig. 4. A f!ine tungsten wire probe which causes only a
negligible shadow makes contact with the illuminated diffused layer. This probe is moved
across this layer along a line perpendicular to the contact strip, while the contact strip
itself is shorted to the back contact of the cell. The voltages, measured with a high-resistance
voltmeter connected as shown in Fig. 4, are then plotted against the corresponding distances
of the probe from the contact strip. Near the terminal strip, the voltage increases practically
linearly with distance from the contact, then increases more slowly and finally, when the
cell open circuit voltage is reached, remains constant. This saturation at large distances
from the contact has two reasons: (a) the density of the current flowing in the diffused
layer toward the contact strip decreases with distance from the contact, and (b) with in-
p:
1
F
L.
5
P
I.
:
:b.
?.
Iv
Series Resistance Effects on Solar Cell Measurements
463
creasing voltages across portions of the cell the current contribution from those portions is
reduced (compare with current-voltage curves such as Fig. 2).
Even if solar cells are operated at light intensities such that the current density in the
diffused layer is lower than required for saturation, high enough voltages may develop in
portions of the cell to cause a substantial reduction in the current contributed from the
corresponding regions of the cell. This mechanism appears to be the only one by which the
cell can maintain short circuit conditions at its terminals, thereby explaining the aforementioned differences between short circuit current and light generated current.
The foregoing discussion emphasizes the caution which has to be used in the application
of the short circuit current of a solar cell for light metering purposes at high light intensities.
The amount of internal series resistance in the cell has to be determined and the region of
deviation from the linearity be established before a cell can be confidently applied for such
purposes.
4: THE
EFFECTS
OF INTERN
so LAR CELL
AL SERIES
P ERFORMANCE
RES ISTANCE
ON
The curves of Fig. 2 indicate that the internal series resistance can severely afiect the
performance of photovoltaic cells as solar energy converters. The maximum power output
of a solar cell is given by the area of the largest rectangle that can be drawn inside the
photovoltaic output characteristic. The area of such a rectangle increases with increasing
“sharpness” of the knee in the photovoltaic output characteristic. Internal series resistance
causes a successively larger “rounding” of the characteristic at increasing light intensities.
Solar cells are normally designed for best performance at radiation intensities as obtained
at the earth’s surface. Such cells may therefore not give optimum performance at increased
light intensities as may be encountered in space vehicle applications in closer proximity to
the sun or in the use of solar cells combined with radiation concentrating devices.
Figure 5 presents three sets of curves, calculated for a solar cell with 10=6.9 x lo-l2 A,
IL = 5.5 x 10-Z A, and q/&T = 38 6 V-1, and with assumed internal series resistance
values of 0, 0.5 and 1.0 Q. The curves give the maximum power output into a matched
load, the maximum power point voltages and the efficiencies for values of solar irradiance
between 50 and 350 mW/cm2. The zero-resistance solar cell shows monotonously increasing
power output with increasing light intensity. Finite values of the series resistance cause
the maximum power output to increase less rapidly at the higher light intensities . This
effect becomes progressively larger with increasing series resistance and light intensity.
The maximum power point voltage, which steadily increases with increasing light intensity for the zero-series resistance cell, starts to decrease at higher light intensities for the
cell with 0.5 Q series resistance, and shows continuous and larger decrease at a series
resistance value of l-0 Q through the intensity range of 50 to 400 mW cm-2. The maximum
power point voltage which steadily increases with light intensity for the zero-series resistance
cell starts to decrease at the higher light intensities for the cell with 0 * 5 Q series resistance,
and shows a continuous drop at a series resistance value of 1.0 Q.
The dotted curves in Fig. 5 present the conversion efficiency. The curve for the zeroseries resistance cell shows a monotonous rise toward the higher solar irradiance values
while the curves for the 0 * 5 and the 1.0 Q cells exhibit progressive efficiency decreases
toward the higher light intensities.
Also presented in Fig. 5 are experimental curves obtained on two solar cells with 0.38
and 3.5 Q series resistance, respectively. It is interesting to note that the shape of the experil
464
MARTIN
and
WOLF
HANS
RAUSCHENBACH
mental curve for the 0.38 Q cell fits perfectly into the system of the theoretical curves
although its absolute values are somewhat lower due to deviations of other device parameters from those assumedfor the theoretical evaluation. The information presented in
this paragraph leads to the conclusion that for high power output and high efficiency at
MAXIMUM
POWER
POINT
VOLTAGE
---EFFICIENCY
i
-1
r
z
70
l
LO
I
1
so
/
I
1
I
----too
SOLAR
IRRADIANCE
1
200
IN MILLI
FIG.
,300
WATT
PER
CM
5.
high light intensities, and for a minimum change of the maximum power point voltage
with increasing light intensities, solar cells with progressively smaller internal series resistance will have to be selectedor be specifically developed.
5. A METHOD
FOR THE
CHARACTERISTIC
PREDICTION
OF THE
FOR DIFFERENT
PHOTOVOLTAIC
LIGHT
LEVELS
OUTPUT
The photovoltaic output characteristic as described by equation (1) permits an easy
prediction of the corresponding current-voltage characteristic for a given light level which
differs from the one at which the original characteristic has been measured.
Series Resistance Effects on Solar Cell Measurements
465
In the case of the solar cell with zero-series resistance, the exponent of equation (1)
contains only the independent variable, that is the terminal voltage. For any fixed light
level only the exponential term is variable on the right-hand side of equation (1). Any
change in the light intensity enters then into the relationship through the light generated
current IL. A change in the light intensity results, at constant terminal voltage V, in a change
of the terminal current I equal to the difference of the light generated currents AIL. This
meansthat the shapeof the curve, which is determined by the exponential term, is invariant
with respect to light intensity. The mathematical derivation for this is presented in Appendix
A of this paper. The photovoltaic output curve for a different value of light intensity can
readily be obtained by translating the original curve parallel to the ordinate of the IV coordinate system by an amount equal to the difference in the light generated currents, which
is proportionate to the difference of the light intensities.
The method outlined in the previous paragraph is valid for solar cells with small values
of seriesresistanceor at sufficiently low light levels so that the effects of the seriesresistance
can be neglected. This is expressedin the condition that the product IRS has to be negligibly
small compared to the terminal voltage V. If this condition is not fulfilled, then a second
translation parallel to the abscissahas to be performed on the photovoltaic output characteristic for the transformation to a second light intensity. The second translation consists
in a decreaseof the terminal voltage by an amount equal to the product of the internal series
resistance Rs and the difference in the light generated currents, AIL. A formal derivation
for this caseis also given in Appendix A.
For a change of the photovoltaic output characteristic from the light intensity L1 with
the light generated current ZL~to the light intensity L2 generating the current ILL, the two
translations
I:! = 11 - AIL
(4)
and
V2 = VI - AILR~
(5)
have therefore to be performed. Since IL is proportional to the light intensity IL = CL,
AIL can be expressedthrough a relationship between the light intensities:
AIL = C(L2 - Ll)
(6)
Since ILL can be readily obtained from the photovoltaic output characteristic, measured at
light intensity L1, AIL is more conveniently related to the difference in light intensities
through the form:
Jk--2
AIL = ILL ~
(7)
Ll
A previously expressednote of caution in the substitution of the short circuit current Isc
for the light generated current IL at high light intensities or large value of seriesresistance
should be observed.
Figure 6 illustrates the two translations of the coordinate system to be performed by
going from a light intensity L1 at which the photovoltaic output characteristic has been
measured,to a higher light intensity L2 at which this characteristic is desired to be known.
Figure 7 presents experimental data to demonstrate the correctness of this procedure.
The photovoltaic output characteristics from Fig. 2 for the low seriesresistance solar cell,
measured at five different light levels between 50 and 400 mW cm-z, are reproduced in
Fig. 7. After performing the translations of the coordinate system for the four photovoltaic
output curves obtained at the lower light intensities to the 400 mW cm-3 heI.
they are
Found to all fill1 on top of each other as well as on the experimental curve for this intensity,
The experimentally obtained curves are given in solid lines, while the transformed curves
are presented by squares. circles, and triangles, corresponding
to the different original
intensities.
AI,*
R,
FIG.
6.
A
IMETHOD
FOR
6.
THE
DETERMINATION
SERIES
RESISTANCE
OF
THE
INTERNAL
An inversion of the method described in section 5 permits the easy and accurate determination of the internal series resistance of any solar cell. For this purpose the photovoltaic
output characteristic has to be measured at two different light intensities, the magnitudes
of which do not have to be known. The two characteristics
are translated against each
other by the amounts AIL and LIIL x R, in the y- and the .x-direction, respectively. Two
corresponding
points on the two characteristics show a displacement with respect to each
other, which is identical to the two translations of the coordinate systems. The displacement
parallel to the ordinate gives the value of AIL. Since the displacement parallel to the abscissa
equals L&R,, the value of the internal resistance Rs is readily obtained.
One practical approach to this procedure is to choose an arbitrary interval 431 from the
short circuit current &, which determines the first characteristic.
It is frequently found
convenient to choose d/ so as to obtain a point in or near the knee of the characteristic.The
same Al value is used for the finding a second corresponding point on the second characteristic. An illustration of the procedure is presented in Fig. 8.
This method was first suggested in 1960 by Swanson
7.
EXPERIMENTALLY
FOUND
CONSTANT
DEVIATIONS
MODEL
[7].
FROM
THE
LUMPED
It has been found that the measured current-voltage
characteristics
of many solar cells
cannot be very accurately described by the equivalent circuit model of Fig. 1 and by equation (1). These deviations have been found to be due to two separate effects: one involving
the P-H junction characteristic,
and the other the internal series resistance. The effect
k
467
Series Resistance Effects on Solar Cell Measurements
involving the p-n junction characteristic will be discussedin section 8, while in this section
attention will be directed to the effect of the internal series resistance.
!“~“TOVOLTAIC
E
TRANSLATED
0.010
0.038
0.057
FIG.
7. Current-voltage
VOLT
VOLT
VOLT
TERMINAL
VOLTAGE
IN VOLTS
BY
50MlLLl
IOOMILLI
I50MILLI
characteristics of solar cell No. A-M-12 at different illumination
levels.
If the model discussedin the previous sections is valid, then the samevalue of internal
series resistance Rs will be obtained by application of the method described in section 6,
independent of the chosen value of Al. Also, a sequenceof photovoltaic output characteristics obtained at different light levels, and transformed to any one of these curves by
the method described in section 5 will show perfect agreement between the individual
curves as illustrated in Fig. 7. Equality of the seriesresistancevalues obtained from different
parts of the characteristic, and agreement between transformed curves is, however, not
found on all solar cells measured. The high seriesresistancecell A-N-16 used for obtaining
MARTIN
E-IANSRAUSCHENIMCH
and
WOLF
I
*L
FIG.
TERMINAL
0
-,.
4-30
0.1
?
0.2
VOLTAGE
0.4
0.3
R,
8.
IN VOLTS
05
SQLAR CELL A-8-37
AVERAGE
SERIES
_ RESISTANCE
R,=0.56il
-
i
J
z
z-40-
---
I5
QT
$50,
2
z
,=5OMA
S-60
e
-70
1
I
Aq= SOMA
I
Ah
=30MA
I
FIG.
9(a)
--_
0.6
0.7
Series Resistance Effects on Solar Cell Measurements
469
data for Figs. 2, 3 and 5 shows such deviations. These observations are not only made on
cells with high internal series resistance, although the deviations found in low series resistance
cells are generally substantially smaller.
VARIATlON
INTENSITY
Of CURVE
OBSERVED
SOLAR
=EE .
IRRADIANCE
MW
I
2
3
I
SHAPE
WITH
LIGHT
ON CELL
A-B-37
SHIFT
CM*
SO
. -
100
2s
SHI~vA*L
4
300
12s
63
5
400
173
90
0.1
0.2
0.3
VOLTAGE
0.4
Rs
13
30
7s
200
TERMINAL
0
AI
MA
IN VOLTS
0.5
0.6
0.7
2
6
2 -10
-
z
z
s-20aw
5
u
$ -3o5
5
k- 40-
FIG.
9(b)
Figure 9(a) presents two curves obtained on a low series resistance gridded solar cell at
two different light intensities. Marked in this figure are the series resistance values obtained
at different portions of the characteristics. These values range from 0.5 to 0.7 Sz. It has
been observed, that the limits of the range of series resistance values are rather independent
of the amount of light intensity and of the difference in light intensities used for this evaluation, except for the fact that at high light intensities a different portion of the curve can be
evaluated which is not available at the low intensities. This is demonstrated by the data
contained in Table 1, which were obtained on cell A-G-43.
470
MARTIN
TABLE
Initial
SOhl-
1. SERIES
Initial light
generated
current
irradinnce
(m Wcrn- “)
-
50
WOLF
and HANS
RESISTANCE
VALUES
ODTAINED
CORRELATION
POINTS AND
Final
solar
irradinnce
(niWcn+)
Final light
generated
current
RAUSCHENI~ACH
ON SOLAR
CELL
LIGIIT
INTENSITIES
Light
generated
current
difference
A-G-43
AT DIFFERENT
Correlation
current
diference
JIr,
.-I I
(t-n A)
(mA)
(imA)
lLZ
kfeasured
series
resistance
25.1
-100
50.1
25
2.7
5.6
12.1
20.2
25. I
0.4
0.4
O-4
0.36
0.36
-loo
50-l
-200
100.0
49.9
10.0
20.0
30.0
40.0
49.9
0.36
0.34
0.36
0.36
0.32
W-100
co* 1
-300
150-o
100
5-O
10-o
25.0
50.0
0.58
0.42
0.37
0.34
f-100
50-l
-400
199
148.9
5.0
25.0
50. I
0.43
0.37
0.34
-100
50.1
-500
249.5
199.4
5.0
25.0
50.1
0.38
0.39
0.36
-300
150.0
-400
199
49
25
150
0.36
0.30
-300
150-O
-500
249.5
99.5
25
50
75
100
125
150
0.42
0.39
0.37
0.36
0.34
0.32
--
-
The reasons for the deviations are to be found in the physical configuration of the solar
cell. Major contributions to the internal series resistance come from the sheet resistance of
the p-layer, the bulk resistance of the n-layer, and the resistance between the semiconducting
material and the metallic contacts. While the contact series resistance can properly be represented by a lumped resistance, the sheet resistance and the bulk resistance are distributed
throughout the device. Figure 10 shows the transition from the physical configuration of
the solar cell to the distributed constants model. The cell is imaginarily divided into many
elements, each one containing a p-n junction, which acts as a source and a shunting diode
simultaneously. These junction elements are inter-connected by the distributed resistances.
This model was studied in 1958 by the first author as the proper representation for the
current flow and voltage distribution in a solar cell, but has been discarded because of the
enormous complexity of its evaluation. A similar, slightly simplified model has been proposed by Wysocki, Loferski and Rappaport in 1960 [8], but was apparently also not
SeriesResistanceEffects on Solar Cell Measurements
evaluated. Because of the rather small values of internal series resistance encountered
the modern gridded solar cells, an intermediate model is proposed here. It represents
next rehnement step from the lumped constants model of Fig. 1. Instead of dividing
cell into many elements, it splits the solar cell into two equal parts and adds the bulk
SOLAR
PHYSICAL
in
the
the
re-
cu
CONFIGURATION
,lLr”i’i^i’gi
,
471
,
~
7POSlllVE
1
ELECTRICAL
CONTACT
NEGATIVE
CONFIGURATION
POSITIVE
CONTACT
RESISTANCE
CONTACT
CONTACT
CONTACT
RESISTANCE
L
DISTRIBUTED
2ND
ORDER
NEGATIVE
CONSTANTS
LUMPED
CONSTANTS
&I
FIG.
CONTACT
MODEL
MODEL
R52
10.
sistance, the contact resistance, and a portion of the sheet resistance into a lumped element
next to the terminal. The other portion of the sheet resistance is placed in the connection
between the two parts of the cell. The ratio between the values of the two resistance elements
has been found to vary between individual cells.
Figure 12 presents the p-n junction characteristic of a silicon solar cell obtained by the
third method described in section 2. It also presents theoretical points for the photovoltaic
output characteristic derived from the p-n junction characteristic by using both the lumped
constants model according to Fig. l(a), and the improved lumped constants model of
Fig. 10. Also presented is the experimentally obtained photovoltaic output characteristic
18
472
MARTIN
WOLF
and HANS RAUSCIIENIIACI~
WIjich demonstrates good agreement between the experimental data and the improved
lulllped constants model. A resistance ratio of 9 to 10 11:~ becl~follnd to be appropriate for
the cell evaluated for this tigurc.
8.
APPLICATION
OF
THE
P-N
FOR
DEVICE
JUNCTlON
STUDIES
CHARACTERISTIC
The diode forward characteristic has long been used for the investigation of device
parameters like the saturation current. the constant J in the exponent of the diode characAMP
SOLAR
CELL
Voltage,
FIG.
A-N-12
volts
11.
teristic, the temperature dependenceof these characteristics, and for the interpretation of
the physical mechanismscausing the observed device behavior. The p-n junction characteristic, described as method 3 in Section 2, provides a very useful supplement to the diode
forward characteristic for such studies. The usefulness of the p-n junction characteristic
stems from the severe effect which the seriesresistance has on the diode forward characteristic at high injection levels, and which is not present in the p-n junction characteristic.
At sufficiently low injection levels where series resistance effects become negligible the
two characteristics become identical. Figure 11 contains a diode forward characteristic and
r
Series Resistance Effects on Solar Cell Measurements
,-
473
a p-n junction characteristic obtained at 23°C on the gridded solar cell A-N-12 which was
used for the previously described measurements.
The graph shows excellent agreement
obtained between the two different methods at injection levels below 10-Z A on the 2 cm2
cell. Experimentally
obtained points of the p---72junction characteristic
given by circles
superimposed on the also experimentally obtained diode forward characteristic, presented
by the solid curve demonstrate this agreement. Above 2 >: 10-Z A, however, divergence
between the two curves is observed. The difference between the ~-II junction characteristic
and the diode forward characteristic
at high injection levels corresponds to a series resistance of 0.33 Q, which differs from the 0.38 Q value obtained from the photovoltaic
output characteristic according to section 6.
Evaluation of the reverse characteristic yields a shunt resistance value of 13.8 kSZ, and
a saturation current 10 = 6.5 x 10-G A. The current conducted through such a shunt
resistance is shown in Fig. 11 by the curve marked “I = V/I&,“. By subtracting the current
conducted through this shunt resistance from the current values of the diode forward
characteristic, a new curve is obtained which deviates from the experimentally obtained
characteristic at the 10~~ injection levels. The curve thus obtained is the true P--J? junction
characteristic without resistive shunt effects. The curve is matched by a diode equation of
the form of equation (2) with a saturation current IO = 5.5 x 10-c A and constant
A y: 2.86 at a device temperature of 296°K. At the high injection levels, however, a deviation from the exponential is observed. This deviation is in the diode forMvard characteristic
partially hidden by the effect of the series resistance, but can be clearly observed and
evaluated from the p-11 junction characteristic.
The p-17 junction characteristic
at high injection levels asymptotically
approaches a
second exponential curve which in the semi-log plot of Fig. 11 is represented by a straight
line. The transition region between the two pure exponential portions of the curve is very
accurately described by the sum of the two exponential terms. It is interesting to note that
on all solar cells thus investigated, the second exponential term has an A value of 1 and
saturation currents of 10-l’ to lo-11 A for the 2 cm2 junction. Such values are to be expected
from a deFTice following the carrier diffusion theory of a p-n junction as described by
Sliockley [9].
Also sho~+~n in Fig. 11 is a diode forward characteristic obtained on the same cell at a
device temperature of 417°K. At the high injection level a series resistance voltage drop
corresponding to the series resistance value of 0.33 Q was found in the room temperature
curve. This voltage drop was subtracted from the diode forward characteristic and, again,
was found to lead to a good match to the p-n junction characteristic. The very highest
current values, which were obtained by the diode forward method only. appear too high.
probably due to thermal runaway during the measurements.
The resultant characteristic for the p-n junction at 417°K can also be best described bj
the super--position of two exponential terms. Like in the room temperature case, the curve
approaches asymptotically
the exponential represented by a straight line in the semi-log
plot at the high injection levels, while the transition region is accurately described b)r the
sum of the two exponentials.
The remarkable finding is, that the exponential for the higher injection levels has the
constant A = 1. Using the carrier diffusion theory for p-/r junctions as described by
Shockley, and the experimental value for IO = 3 * 3 s 10-l” A at 296°K. one arrives at a
saturation current 10 = 8 ?: 1O-G A for the same junction at 417°K. This is in very good
agreement with the experimental finding of 7 x 10dGA.
ji
I
‘I
:( i
i”
i;
I .
474
MAkTrN
WOLF
and
HANS
RAUSCHENBAC~I
While extremely good correlation between the p-n junction characteristic at high injection levels and the carrier diffusion theory for p-n junction behavior could repeatedly be
established, no correlation to a theoretical behavior could be found for the lower injection
SOLAR
-2
CELL
A-G
-43
-5
F’HOTOVOLTAIC
---.--7
g
<
OUTPUT
CHARACTERISTIC
EXPERIMENTAL
CURVE
o o o
THEORETICAL
POINTS,
rc.w( v- f 1 cJ.36)
1=1.6x KY [ lo
IO
-
00
THEORETICAL
POINTS.
LUMED
IMPROVED
CONSTANTS
MOD.
MODEL
7
-1
ii
-12
z
P-N
c
z
aw
Q -I5
--V
s
JUNCTION
CHARACTERISTIC
EXPERIMENTAL
V THEORETICAL
CURVE
POINTS,TWO
EXPONENTIALS
I =1.6~16’%0’~~~‘“-l)+3x10~
( IO+~“-I~
(AMP)
-17
-20
SINGLE
EXPONENTIAL
1 =l.~xI(y
( lo’6.9’
-\
“-I)
-22
-25
FIG.
12.
level part of the characteristic. A considerable spread of values for the constant A and for
the saturation current were found for the lower injection level exponential. Still more nonuniformity was observed in the temperature dependence of this portion of the curve. Since
no pattern could yet be established in these findings, a hypothesis shall not be advanced
for the physical nature of the characteristics at the lower current levels.
The foregoing discussionslead to a replacement of equation (1) by the following form :
I=lol
[exp(-&f)-
l]
$-lo2
[exp(&V)
with A = A(T), excepting the seriesresistance terms.
-
l]
-IL
Series Resistance Effects on Solar Cell Measurements
475
According to this description a new equivalent circuit diagram is evolved containing two
separate forward biased diodes connected in parallel. This new model resembles one
advanced earlier by Watson [lo]. Watson, however, assignedidentical exponents to both
junctions, and described only a relatively small difference in the saturation currents. A
separation of the two p-n junctions by the series resistance is not necessary, because the
p-n junction characteristic alone exhibits the property of the super-position of the exponentials, and at high injection levels where the seriesresistancebecomesmost effective the
characteristic is dominated by only one of the two exponentials.
I
I
D, = 10,
-+[V-IRsr
IO2 = b++&[V-1R.r
IO3
=Io,(EXR
-(~-+f~+I~~+lo~)
(I-
c
5
+-I‘+
Rr;1
*.,+I.,,.,,]]
-I}
A= f (1)
(V-IRs#$
FIG.
13.
The knee of the I-V characteristic, however, is most heavily affected by the transition
region between the two exponentials. This is demonstrated in Fig. 12, which shows an
exponential curve with A = 1 and 10 = 1.6 x 10-12A together with the experimentally
obtained p-n junction characteristic for cell A-G-43. In the portion of the characteristic
nearer the open circuit voltage point, which is the high injection region, close approximation between the two curves can be noticed. In the knee of the curve, however which falls
into the transition between the two exponentials, the experimental curve appears much
more rounded than the single exponential with A = 1. In the part of the curve approaching
the short circuit current, the difference between the two exponentials is not discernible
becauseof the scale of the plot.
The sum of two exponentials, for which the saturation currents and A-values have been
obtained from a semi-log plot for cell A-G-43 similar to Fig. 11, has been calculated, and
its results are superimposed on the experimentally obtained p-n junction characteristic,
illustrating the good agreement.
Finally, the photovoltaic output characteristic is shown, obtained experimentally at a
light intensity equivalent to about 500 mW cm-2 of solar irradiance. Shown with this curve
as circles are values calculated by using the “sum-of-two-exponentials” model together
476
MARTIN
WOLF
and HANS RAUSCHENBACH
with a single lumped series resistance corresponding to the original model of Fig. I(a).
The points obtained in this manner show a fair approximation to the experimental curve,
with the largest deviations at the knee of the curve which is of greatest interest for maximum
power output considerations.
A further calculation has been carried out using the improved lumped constants model
described in section 7 and Fig. 10 together with the “sum-of-two-exponentials” model.
The results are presented in the form of squares superimposed on the experimental curve.
They prove the very close approximation to the actual cell characteristic obtained by the
use of this model. The equivalent circuit diagram of this model is shown in Fig. 13, and the
equation describing it has the following form:
I = loI exp -&(V(
[
i- 10~
(
exp
[
IR.92)] - 1) + IO2(exp [-&,(v
- IRA]
- 1)
11
-
~[V-IRsz-(I--tlr,+ID,+Io,)R,,]
1
Although this model is only the next improvement step after the original lumped constants
model, it exhibits already considerable complexity, but yields also much more satisfactory
results.
It will be interesting to note that for cell A-G-43 the values for Rs, and R,, which were
found to give best results were O-265 and 0.30 respectively. This compares to a single
lumped resistance of Rs = 0.36 Q for the old model. The values indicate a larger contribution from the sheet resistance than from the bulk and contact resistances.
The foregoing discussions shed light on the importance of the p-rz junction characteristic to the device development engineer, who can by applying this method study the
device parameters without being hindered by the overshadowing effects of seriesresistance.
9. CONCLUSION
The foregoing discussionilluminates the effect of seriesresistanceon solar cell measurements and the caution to be taken in their interpretation. The reduction of efficiency at
high light intensities due to high values of seriesresistancewas analyzed with the conclusion
that for advantageous application of solar cells at high light intensities emphasishas to be
placed on the reduction of series resistance. Finally, new features of the p-n junction
characteristic were discussed.
These were previously hidden due to the modification of the current-voltage characteristic by series resistance effects. New lumped constants models were developed which
more closely approximate the characteristics measured on the solar cells.
Acknowledgement-Special
appreciation shall be expressed to Messrs. P. Goldsmith and W. Schaefle of
the Jet Propulsion Laboratory, Pasadena, California, for their mentioning of observed discrepancies between
measured photo-voltaic output characteristics and the theoretical curves as described by the older lumped
constants model.
477
Series Resistance Effects on Solar Cell Measurements
APPENDIX
A
Finding the proper IV characteristic for an individual solar cell or a solar cell matrix
after a light level change by meansto two translations of the coordinate system
The basic equation describing the current-voltage characteristic of a solar cell neglecting
series- and shunt-resistance is :
I = 10(eBv - 1) - 1~
tfw
where: B = q/AM’, and all other quantities as described in section 2 of the paper.
Let :
11 = IO (eBVI - 1) - 1~~
CA9
be equation (Al) at light level L1 and
I2 = IO (eBv2 - 1) - ILL
643)
the same equation at light level L2.
Since V is the independent variable, one can choose:
v2 = Vl
w9
and can set:
1~~
=
1~~
+
W)
AIL
where AIL is proportional to the difference in light intensity between levels 1 and 2. Subtracting equation (A2) from equation (A3) after introducing equation (A4) and (A5), one
obtains :
I2 = I1 - AIL
(fw
for all choices of I$ = IQ.
Equation (A6) describesa translation of the coordinate system parallel to the current
axis by the amount AIL on the current axis.
For higher light levels, the effect of seriesresistance on the IV characteristic has to be
included, due to the increased magnitude of the current I (see Fig. 1). Here
V’ = V - IRS
(A3
is the voltage across the p-n junction, which is larger than the terminal voltage V by the
voltage drop in the seriesresistance. (Note that the current I is a negative quantity (see
equation (Al)), resulting in V’> V for power generation in the solar cell (4th quadrant
operation). The 1Y characteristic for the equivalent circuit (Fig. l(a)) is:
I = IO(eB(V-1Rs) - 1) - IL
= Il(eBV’ - 1) -IL
W)
(A8a)
Introducing again two light levels 1 and 2, one obtains:
11= I1 (eBv$ - 1) - 1~~
12= I1 (eBV’2 - 1) -
1~~
-
(4
AIL
w0)
Again one chooses:
74 = V’l
and obtains the same translation as before:
12 = 11- AIL
(Al 1)
W)
478
MARTIN
WOLF
and
HANS
RAUSCHENBACH
Equation (A7) results, however, in two different terminal voltages Vr and Vz for the two
currents 12 and Il. From equations (Al l), (A7) and (A6), follows:
VI - IlRs = V2 - IlRs + AILR,
(Al la)
which describes a constant relationship between VI and V2 for any choice of VI. The
constant of this relationship is proportional to the series resistance Rs and to the change
in light level. Equation (Al 1a) thus describes a second translation of the coordinate system,
this one parallel to the voltage axis by the amount:
V, = VI - AILR~
@q
Equations (A6) and (A7) describe the photovoltaic output characteristic as invariant to
any change in the light intensity except for two translations of the (current-voltage) coordinate system.
REFERENCES
[l] H. RAUSCHENBACH,
Understanding Solar Measurements, Semicond. Prod. Applic. News. Hoffman
Electronics Corp., El Monte, Calif. (1960).
[2] W. G. PFANN and W. VAN ROOSBROECK,
J. Appl. Phys. 25, 1422 (1954).
[3] M. B. PRINCE, J. Appl. Phys. 26, 534-540 (1955).
[4] M. WOLF and M. B. PRINCE, “New Developments in Silicon Photovoltaic Devices and Their Application in Electronics”, plublished in Solid State Phys. Electron. Telecom., Proc. Int. Conf. Brussels, Belgium,
June 2-7, 1958,2, 1180-l 196, Academic Press, New York (1960).
[S] A. J. HEEGER and T. R. NISBET, The Solar Cell Diode Curve, Rept. LMSD No. 310003, Lockheed
Missile Systems Division, Sunnyvale, California (1958).
[6] H. J. QUEISSJZR, SoZid State Electron. 5, 1 (1962).
[7] L. D. SWANSON,
private communication.
[8] J. J. WYSOCKI,
J. J. LOFERSIUand P. RAPPAPORT,
Research on Photovoltaic Converters, Proc. 14th Ann.
Power Sources Conf., pp. 32-36, Power Sources Div., U.S. Army Signal Res. and Dev. Lab., Ft. Monmouth, N.J. (1960).
[9] W. SHOCKLEY,
EZectrons and Holes in Semiconductors, Van Nostrand, New York (1950).
[lo] R. H. WATSON, Equivalent-circuit Characteristics of the Solar Cell. Presented at the 1960 Pacific General
Meeting of the AIEE at San Diego, California, August 11, 1960.
Zusammenfassung-Strom-Spannungs-Charakteristiken
von photoelektrischen
Sonnen-energie-umwandlem kann man nach drei verschiedenen Methoden erhalten. Durch den Einfluss des Innenwiderstandes der
Zellen liefert jede ein anderes Ergebnis. Die drei resultierenden Charakteristiken sind: (1) Die photoelektrische Ausgangs-Charakteristik,
(2) die p-n-fIbergang-Charakteristik
und (3) die Gleichrichter-DurchlassCharakteristik. Urn die fur die individuelle Anwendung richtige Information zu erhalten, muss in jedem
Fall die geeignete Methode ausgewghlt werden.
In den meisten Fallen wird die photoelektrische
Ausgangs-Charakteristic
benutzt, z.B. fur die Bestimmung der Leistung von Sonnenenergie-Umwandlern.
Es wird eine Methode beschrieben, die es gestattet,
eine solche Charakteristick fur eine beliebige Lichtintensittit aus einer entsprechenden anderen mit davon
verschiedener Lichtintensitat rasch herzuleiten. Die Methode benutzt zwei Koordinaten-Transformationen
und erfordert nur die Kenntnis des Innenwiderstandes und die Abhangigkeit der Kurzschluss-Strome von
der Intensitat.
Eine Umkehrung dieser Methode erlaubt eine einfache Bestimmung des Innenwiderstandes. Dazu
genugen Messungen bei zwei beliebigen Intensitaten unbekannter Grosse.
Die Wirkung des Innenwiderstandes bei hohen Lichtintensitaten besteht in einer Abflachung der photoelektrischen Ausgangs-Charakteristic
und damit einem Abfall der Spannung bei der grossten Leistung.
Die daraus resultierende Verkleinerung des Wirkungsgrades muss bei Anwendung von optischen Konzentratoren oder bei Verwendung im Weltraum in grijsserer Sonnennghe durch Erniedrigung des Innenwiderstandes ausgeglichen werden. In Zellenbereichen grosseren Abstands vom Kontaktstreifen w&h.st die
Spannung an und erreicht bei grossen Intensitaten fast die Verhaltnisse beim offenen Stromkreis. Das
bedingt eine Stromverdrangung von diesen Bereichen der Zelle und eine Abweichung von der normalen
Proportionalit&
von Kurzschluss-Stom und Lichtintensitgt.
Die direkte Messbarkeit der p-n-Ubergang-Charakteristik
nach der zweiten Methode ohne Innenwider-
Series Resistance Effects on Solar Cell Measurements
479
standseffekte bei hohen Stromdichten erweist Sikh als hervorragendes Werkzeug fiir den Entwicklungsingenieur und liefert daneben noch eine zweite Miiglichkeit fiir die Bestimmung des Innenwiderstandes.
Die Ergebnisse der Anwendung dieser Methode zeigen, dass die Charakteristik der Silizium-Photozelle
im Stromdichte-Bereich der Sonnenbatterien durch die Diffusionstheorie fiir p-n-uberggnge
weit besser
beschrieben wird, als man friiher angenommen hatte.
R&urn&-Les
CaractCristiques Intensit&Tension
des cellules de convertisseurs d’energie solaire photovoltaiques peuvant &re obtenues par trois m&hodes qui donnent des rCsultats diffkrents par suite des effets
de r&stance intCrieure en sCrie des cellules. Les trois caractkristiques rCsultantes sont: (1) la CaractCristique
de rendement photovoltaique, (2) la caractCristique de jonction p-n, et (3) la caractCristique de redresseur
(vers l’avant). Pour obtenir le renseignement correct relatif i l’application particuli&re, il est nCcessaire de
choisir la mCthode approprike.
La caractCristique de rendement photovoltaique est utilisCe t&s frlquemment, par exemple, pour la
dktermination de la performance du convertisseur solaire. 11 est d&it un pro&d6 rapide permettant de tirer
cette caractkristique pour tout niveau de lumi&e d’une caractkristique correspondante obtenue pour un
niveau de lumi&re diff&-ent. Cette mCthode fait inttrvenir deux translations du sysdme de coordonn&s et
n’exige que la connaissance de la rlsistance en sCrie et de la diffkrence des intensites lumineuses des courants
de court-circuits.
Une inversion de cette mCthode permet de dCterminer facilement la rksistance en sCrie, en faisant intervenir des mesures & deux niveaux de lumibres arbitraires de grandeur non connue.
A des niveaux lumineux Clevls les effets de la rCsistance en sCrie se traduisent par un aplatissement de
la caract&istique de rendement photovoltaique et en une chute de la tension du point de puissance maximum.
La diminution de rendement rCsultante doit etre neutralisee par reduction des rCsistances en sCrie pour des
applications de cellules solaires avec moyens de concentration optiques, ou bien pour des missions spatiales
B proximitt? du soleil. Dans les parties des cellules, progressivement plus CloignCes de la bande de contact, il
se produit des tensions de cellules croissantes se rapprochant de 1’Ctat de circuit ouvert B des intensitis
lumineuses t&s ClevCes,ce qui donne lieu d une rCduction de l’apport de courant provenant de ces
parties de cellules et B une dCviation de la proportionalit
normale entre le courant de court-circuit et l’intensit6 lumineuse. La mesurabilitk directe par la deuxieme mCthode de la caractkristique de jonction p-n A des
densit& de courant Clev&s saris effets de rCsistance en sCrie met g la disposition de l’ingenieur de mise au
point un instrument puissant et, en outre, procure une deuxiCme mCthode de dktermination de la Gsistance
en sCrie. Les rCsultats de l’application de cette mCthode indiquent que dans la gamme des densit& de courants
utiliskes dans la conversion de 1’Cnergie solaire, la caractCristique des cellules solaires au silicium est d&rite
par la thCorie de la diffusion pour des jonctions n-p avec beaucoup plus de pricision qu’on ne l’avait pen&
auparavant.